The Egocentric Predicament

Thomas Nagel

  • The False Prison: A Study of the Development of Wittgenstein’s Philosophy, Vol. II by David Pears
    Oxford, 355 pp, £29.50, November 1988, ISBN 0 19 824487 8

When I was an undergraduate at Cornell in the Fifties, it was the only American university where Wittgenstein’s later work was the object of intensive study. He had died in 1951 and Philosophical Investigations was published in 1953. I remember in those pre-xerox days sharing with some fellow students the typing of The Blue Book and The Brown Book in multiple carbons, from a set available in Ithaca – pre-Investigations texts dating from the mid-Thirties which circulated in samizdat until they were finally published in 1960 as part of the still continuing stream of volumes from the Nachlass.[1]

The atmosphere surrounding the study of Wittgenstein was both thrilling and stifling: philosophy was agony, and it was necessary to immerse yourself in problems so deep you could hardly breathe. Superficiality was the great danger; nothing could be achieved without struggle, either in approaching the problems or in understanding what Wittgenstein said. Above all, there was the sense that it was almost impossibly difficult to express the truth – witness Wittgenstein’s own failure to publish all but a fraction of the huge volume of material he wrote after returning to philosophy at the age of 40, so that, except for the Investigations, he stands in a peculiar relation of diminished responsibility to it.

I still think this attitude is basically right. There is no way to approach Wittgenstein except by getting mired up to your ears in apparently insoluble philosophical problems, and then seeing whether the places where he suggests you put your feet actually enable you to walk. Today Wittgenstein’s name is dropped everywhere as the symbol of an easygoing conceptual relativism, and references to the Private Language Argument and Forms of Life are almost as common as references to the Heisenberg Indeterminacy Principle. But that is a cultural curiosity: Wittgenstein’s work is scarcely more accessible now than it was thirty-five years ago. It is too easy to embrace the solutions without understanding the problems. As David Pears observes, ‘when we read one of Wittgenstein’s discussions of philosophical illusions, there are two things which we may not find it easy to hold together in our minds simultaneously, his success in dispelling them and the depth and difficulty of the problems that produced them.’ In this second volume of a formidable two-volume study, Pears tries to provide access to Wittgenstein’s later work through the deep traditional problems to which it is a response – the response of trying to find a way out. This means that the book is difficult. Pears is not an agoniser either temperamentally or stylistically, but he is true to his subject, and while there are things to disagree with, those who work through it will gain a sense of Wittgenstein’s depth and radicalism.

In line with his title, Pears traces the development of Wittgenstein’s thought largely in relation to a theme that has been central in Western philosophy since Descartes, and which Wittgenstein found in Schopenhauer: the suspicion that we are trapped inside our own minds and that nothing we can do in the way of language, thought, imagination or perception will enable us to reach beyond them. Descartes thought of this as a problem about what we can know. He assumed that we could at least form the conception of an external world: the issue was whether we could know anything about it or not. But in the development of modern philosophy through Kant, this evolved into an even more radical doubt about what we can think. Even if there is a world beyond our minds, there seems no way for anything in our experience to make contact with it or represent it as it is in itself, so that the reach of our thought is limited to our own actual or possible experiences, including the experiences of ‘external’ perception.

This position is unstable, however. If even our thoughts cannot reach beyond our minds, the idea of an unreachable world beyond is a conceptual illusion: no such thought is possible; anything we take for the thought of what is unthinkable by us must be something else, or gibberish. We cannot say that we can think of the world only as it appears to us, because the implied contrast is meaningless.

In Wittgenstein’s first book, the Tractates Logico-Philosophicus (1921), this result is embodied in the position that solipsism coincides with pure realism. Everything in the world is equally real – from my sense impressions to the stars – but still the world is my world. This shows itself in the fact that however objectively I describe the world (including gaps to be filled in by the things that I don’t know, and even if that objective description affords no privileged position to the particular person I am), I can always add redundantly: ‘And it is I who am saying and thinking this.’ ‘Everything in the world’ is an expression of my language. Yet I cannot think that I am trapped, for that would require, impossibly, forming the thought of what was unthinkable by me.

The Tractatus offered a general theory of language and thought, and of their relation to the rest of reality. Wittgenstein’s later writings reject the possibility of any such theory, in favour of piecemeal description of samples from the great variety of linguistic and mental activities of which humans are capable. But the egocentric predicament remains a central occupation. Wittgenstein continually raises and attempts to dissolve doubt about the adequacy of our language and thought to reach the world it aims to grasp. The details of his response depend on the particular kind of thought he is discussing – about the physical world, other minds, mathematics, action, sensation, meaning. But the general strategy is to make clear an unintelligibility in such doubts that is quite simple: if they are expressible, they must be wrong. If I say, ‘How is it possible that by merely saying the word “Aristotle” I should be able to refer to Aristotle?’ I seem to have forgotten the beginning of the sentence by the time I get to the end – since the second occurrence of the name has to refer to Aristotle for the question to have any sense. I can’t grab hold of Aristotle himself by some kind of super-linguistic reference, against which it is then possible to criticise the ordinary natural conditions of application of the name as inadequate to their purpose.

The same problem of intelligibility arises if I ask how by using the word ‘plus’ I can capture the operation of addition, which is a function defined over all of the infinite possible pairs of numbers, only a small sample of which I shall ever encounter – or the question of whether my ascription of pain to other people really picks out pain, as opposed to some other sensation, or nothing at all. How are the second halves of these questions supposed to be understood, if the questions are serious?

According to Wittgenstein, the philosophical problem arises in each case when we wrench language, mental pictures or other vehicles of thought apart from the conditions which give them significance – as if we could really grasp reality only by transcending those limits. We then use those transcendent ‘thoughts’ to call into question the adequacy of mundane thoughts with their mundane conditions to express what they purport to express. Our actual thought and language come to seem deficient by reference to an unreachable ideal. Pears quotes a passage from ‘Notes for Lectures on Private Experience and Sense Data’,[2] which ends: ‘Isn’t what you reproach me of [sic] as though you said: “In your language you’re only speaking!” ’ Pears adds, ‘Language cannot very well ingest its subjectmatter,’ and: ‘If we think of it as something sublime, spoken from nowhere, the demands that it makes on the world will be too exorbitant to be satisfied and the outcome will be scepticism.’

Much of Pears’s book is devoted to analysis of the development from 1929 on of Wittgenstein’s critique of scepticism about the possibility of conceiving of any mind other than one’s own (a kind of logical as opposed to epistemological solipsism), and less radically, about the possibility of ascribing sensations like one’s own to other subjects of experience on the basis of the merely external observations of those subjects to which one is restricted (the apparent essential privacy of sensation language).

To understand Wittgenstein it is necessary to be able to take these problems seriously. You begin by trying to understand your own system of language and thought from within. You are pulled toward the conclusion that in that system other ‘persons’ are merely objects of your external perception, and that inner experiences, in the sense in which you have them yourself, cannot intelligibly be ascribed to those others. You can’t imagine feeling someone else’s pain, since if you did it wouldn’t be his but yours. So in application to others, the idea of pain is not really that of a feeling but of an externally observable syndrome, composed of circumstances and behaviour. Your idea of real feelings, and of a real self, cannot intelligibly be applied to any but your own case. Use of the same words about ‘others’ is a kind of pun. This is the concept of mind that Wittgenstein was fighting against in the Thirties.

Pears shows how the attack on solipsism prefigures the later attack on private language. In both cases Wittgenstein turns the tables on the sceptic by arguing that the thoughts each of us has about his own case – ‘I’ or ‘The way red looks to me’ – which in the sceptical thought are supposed to provide the secure starting-point from which we try and fail to generalise, become unintelligible if we try to understand them in isolation from their second and third-person analogues: i.e. other people’s identification of a person and of the character of his experiences.

What could be more perfectly secure and transparent, one might think, than my idea of myself – of my ego – and my use of the word ‘I’ to designate it? If I feel a pain in my hand, I know immediately and without need for observation of any kind that it is I who feel it: so my subjective identification of myself is entirely self-contained and independent of the identification of my body or anything else in the world. But Wittgenstein argues that this is an illusion, produced by abstracting language from those connections between experience and its external manifestations that allow us to ascribe experience to one another, and therefore allow us to learn to ascribe it to ourselves.

The possibility of the thought of myself depends on familiar natural facts. As Pears puts it, ‘it is a familiar fact that, when a person’s right hand is hurt, there is a line running into the seat of his consciousness and out again to his mouth. He says “I am in pain,” and though the word “I”, as used by him, does not mean “this body”, it does presuppose the integrity of this personal line. There has to be a connection running back from the mouth that speaks through the seat of the consciousness of pain to the injured part of the body.’ And my infallibility in identifying myself as the subject of my pain – which seems to show a pure ego independent of the body – can in fact be explained, in Wittgenstein’s words, by this: ‘The man who cries out with pain, or says that he has pain, doesn’t choose the mouth which says it.’ My infallible identification of myself as the subject of an experience is merely a linguistic extension of this, rather than the identification of the ‘ego’. It is not, in other words, the application of a purely first-person concept at all.

Wittgenstein was not a behaviourist, as Pears makes clear, but he believed that what Pears calls the ‘personal line’ plays an indispensable role as a condition for the possibility of mental concepts and mental language, and that we cannot use that language to describe circumstances too radically removed from those conditions. This is true not only of the idea of how red looks, or pain feels, to me. He opposed what Pears calls the ‘Dogma of Far-Reaching Sense’, which leads us to think that words carry their meaning with them anywhere.

Passing from the self to the contents of experience, let us consider the irresistible philosophical problem of the inverted spectrum: even if solipsism is false, how can I have any evidence whatever that colours appear to others as they appear to me? Since I can never look into their minds, I can never have any reason to believe that blood doesn’t look to someone else the way grass looks to me, and vice versa – even if there is no behavioural or anatomical difference between us. But the intelligibility of this hypothesis seems to depend on some alternative basis of significance for the ascription of sensory qualities to others, which survives intact even if the usual connections with anatomy, circumstances, and discriminatory and linguistic capacities are imagined to be evidently unusable. And what can this be?

Clearly it is supposed to be a direct projection into the mind of the other of the immediate connection between the word ‘red’ and a particular sensory quality that I experience in my own case, without having to rely on any of these external connections. I stare at the ketchup bottle and ask the simple question: ‘Is what he has or is it not the same as this?

We should be able to see Wittgenstein’s answer coming. ‘The same as this’ is an attempt to stretch an ordinary piece of language beyond the ordinary conditions of its application – to express a super-comparison by reference to which the usual criteria of comparison can be seen as inadequate. But compelling as it seems, it hasn’t been given an alternative basis of significance. As Wittgenstein says, ‘when it is said “Either he has this experience, or not” – what primarily occurs to us is a picture which by itself seems to make the sense of the expressions unmistakable: “Now you know what is in question” – we should like to say. And that is precisely what it does not tell him.’

Not only mat, but even your ascription of sensory qualities to your own experiences, which seem from the inside completely independent of the outward connections to circumstances, behaviour, and standard objects of perception, are according to Wittgenstein dependent for their significance on the existence of such connections in general. If you say, ‘By “red” I mean “the same as this”,’ you are assuming that your subjective idea of a sensation-type is well-defined independently of its connection with anything else. But in fact your first-person identification of colour-impressions as the same or different was learned in connection with capacities for discrimination that show up first in relation to external standard objects, and that can be identified by other speakers who teach you the language.

You may imagine that this is only a device which sets up a connection between words and sensation-types in your mind, a connection in which those external relations have no part. They reply to this proposal brings us to the Private Language Argument and the much-debated subject of rules.

In order for a word to be more than a noise or a scribble or a flicker in the mind – in order for it to mean something – there must be a distinction between its correct and its incorrect application. What is it that determines that distinction? It cannot be whatever the speaker says, for then he wouldn’t be saying anything. The word must carry something else along with it. What is that, and how does the speaker make a connection with it when he uses the word with that meaning?

This is an enormously difficult question. Wittgenstein saw that the background conditions of meaning become invisible to us because words themselves seem drenched with meaning when we use them all the time. So it is difficult even to understand the question, ‘In virtue of what does my word “red” mean a particular type of sensory impression?’ One wants to reply: ‘In virtue of its meaning “red”.’ But that simply repeats what we want explained, as much as if we had said: ‘In virtue of its meaning this type of colour impression.’

Wittgenstein does not offer an analysis of meaning, but he does try to identify certain attempts to use language which cause it definitely to lose its moorings, so that the distinction between correct and incorrect application of the term in question no longer exists, and all we have left is the word (usually in italics). He believed that the mythical private language of experience, completely detached from anything in our behaviour and circumstances that could be observed by others, which is used to express radical scepticism about the experience of others, is an example of this. We fail to see it because we unconsciously rely on our understanding of the experiential concepts of the public language, whose conditions of meaning are concealed in their subjective application to ourselves.

If no one else could in principle tell whether I was using the word ‘red’ (as a term for a type of visual impression) in the same way on two occasions, then neither could I, and there would be no distinction between my using it in the same way and my using it differently. To make sense of the distinction it is not enough, according to Wittgenstein, simply to use the word over again, saying: ‘It’s the difference between applying it to an impression that really is red and applying it to one which isn’t.’ Of course, if there is a difference that is how we will describe it. But something more has to be said about what that distinction might conceivably amount to.

The problem in these cases is that we tend to employ a picture whose usual application is blocked here: we imagine two colour patches side by side, looking different, for example. But what does this image mean in application to the case where I apply the same word to two temporally-separated colour impressions, and there is no conceivable way of any kind to compare them apart from my conviction: in particular, none of the evidence we could usually employ from the public domain – the objects I was looking at, my physical condition, and the correlation between my general identification of colour impressions and other people’s? To raise these questions is not simply to assume a verificationist position, for particular unverifiable claims could be shown to have sense in this way. But it is to require that a type of picture or expression have some conceivable basis of application if we are to believe that the sense of understanding which it gives us is not illusory.

Whether my application of a word in a particular case is correct or not depends on what I mean by it. But what I mean by it is also something about me. So if we are to make sense of the possibility of correcting what I actually say by reference to what I ought to say in light of what I mean, we have to find a comparison between my actual convinced use and something else about me – about what in some larger sense I am doing. It is this crucial distance that vanishes if the application of my language to my experience is imagined logically isolated from everything else as in the mythical private language. We are left with nothing but my successive applications of a term, without anything to compare them with which can count as the rule for its application – the criterion which determines whether I have applied it correctly in each case.

The question of what Wittgenstein thought supplied this need in natural language is a central issue of Wittgenstein interpretation, stirred up recently by Saul Kripke’s proposals and the response to them – which Pears now joins. Pears explains Wittgenstein’s argument that the rule for the application of a term cannot be captured by an ‘instant mental talisman’, for no such thing could by itself determine the distinction between correct and incorrect for an indefinite series of possible future applications: since any such talisman could be interpreted in more than one way, we would still have to explain what makes one of them correct.

All we can do along these lines, in Wittgenstein’s view, is to describe possible grounds of correction for an individual’s use of a term – grounds which the individual himself might recognise as valid. According to Pears, there are two main types of grounds: comparison with the use of others, and calibration on standard external objects of perception which can be re-identified independently of the subject’s application of the term in question. In emphasising the importance of standard objects, Pears agrees with Colin McGinn.

The possibility that this basis for correction might be sufficient for the existence of a rule suggests that a language which was private in the sense that it was invented and used entirely by a solitary being would not be logically impossible: but it would not be logically private, either, since others could in principle learn it, and if they did, comparison with their usage would become a further check on whether the inventor was continuing to use the terms of the language correctly.

If an account of this kind is correct, then we see that the final appeal for the correctness or incorrectness of a particular use of a term is simply more use, of that and other terms, by the user and other persons – not something of a fundamentally different kind. If the challenges we are able to understand have run out, and we have answered every actual challenge by comparisons which do not themselves give rise to further challenge, we are entitled to regard the application as correct. But there is no ultimate, Platonic standard available by which to validate all the judgments to which we appeal in confirming the original one. Beyond a certain point we obey the rules of our language ‘blindly’, as Wittgenstein says, and if we did not, the idea of the valid use of a term would be impossible. It must rest ultimately, if not in the first instance, on judgments we cannot help making. As Pears says, ‘my obedience is “blind” not because I shut out considerations that might have influenced me ... but because, when I have worked my way down to the foundations ... there are no more considerations, doubts, or justifications. I do not even have to listen to the rule, because it speaks through my application of it.’

One issue raised by Kripke’s discussion is whether this position – that the only standard of judgment is a wider circle of judgment – is a form of scepticism. In describing Wittgenstein’s position as a ‘sceptical solution’ to the problem of rules, logic and meaning, Kripke did not mean that Wittgenstein himself would accept such a designation. As Pears says, Wittgenstein thought that scepticism is produced by the demand for a superior standard of correctness of an impossible, Platonic kind, whereas the naturalistic standard of what we and others cannot help doing is all we can have and all we need.

This does not, as Pears thinks, settle the question. We can call someone a sceptic who wouldn’t accept the description himself, if we believe that his account of the basis of a form of thought does not supply an adequate foundation for its actual claims, which reach farther than that basis would warrant. And we may say this whether we are able to supply an alternative basis or not. Thus someone who thinks ethics claims a certain kind of objectivity will regard emotivism as a sceptical theory about ethics, even if it is presented as an analysis of ordinary ethical statements, and even if he cannot himself offer an alternative foundation for the objectivity of ethics. The same might be said of phenomenalism as an analysis of statements about the external world.

The issue in this case is whether Wittgenstein’s view that there is no content to the idea of correctness in the application of any universal term beyond use, broadly conceived, undermines the full strength of the claim to objectivity implicit in our understanding of such terms and what we take ourselves, to be doing when we think with them. I believe that it does, that if Wittgenstein is right, our language is not what we thought it was. What we had thought was that, by picking up on a shared use, we could come to mean a particular arithmetical function by ‘plus’ or a particular colour by ‘red’, and that this ‘meaning’, whatever it is and however we manage it, determines the difference between correct and incorrect application of the term over a range of possible cases infinitely beyond all uses to which the term will be put by us or anyone else, and determines it in a way that is independent of all actual future uses. In other words, meaning reaches vastly beyond conceivable conditions of application, and the limited criteria by which we check use merely provide evidence for it.

Wittgenstein’s reply is that this thought is only apparently intelligible, and that his account does not undermine anything possible, let alone anything real. But even if he is right in this, the issue whether his solution is sceptical depends on whether this illusion of far-reaching sense is internal to our language, or results from imposing on it an unreal external standard which is a philosophical artefact. One might ask what evidence there could be that our language makes stronger claims than it can sustain. This raises the question whether the apparent naturalness of sceptical problems in philosophy itself supplies such evidence. The dissatisfaction and sense of diminishment which Wittgenstein’s arguments often produce in those who find them compelling provides some evidence for the pessimistic view. But Pears, in a striking image, says: ‘they are like people who have been hypnotised and told that they are not standing on firm ground but on a narrow foot-bridge across a gorge, and then the only way to get them to walk is to tell them that there are high parapets on each side of them.’

At the end of the book Pears expresses some sympathy with the feeling that in application to logic and mathematics, Wittgenstein’s account of rules is paradoxical. I see no difference between the cases of logic and experience. That the supposedly infinite applicability of a descriptive term should rest on nothing more than finite agreement in use is just as paradoxical as the dependence of the infinite expansion of a mathematical series on something similar. Our belief in the determinate reality of the infinite expansion or the endless range of cases seems to get no support from the natural human phenomena which Wittgenstein argues so persuasively are all we have got. Pears remarks on the parallel, but I think he errs in finding the position in philosophy of mind and general philosophy of language easier to swallow than the comparable position in philosophy of mathematics.

Here as in other works, Pears’s philosophical writing suffers from intermittent obscurity and an air of impatience. One sometimes feels that steps of briefly presented arguments or answers to questions are left to be supplied as an exercise for the reader. But the book is a real achievement. There is a great deal for the specialist, including criticism and appreciation of other interpretations and detailed historical tracking of arguments and themes through the many drafts and lectures Wittgenstein produced. There is a commendable emphasis on the continuity over a lifetime of Wittgenstein’s underlying philosophical concerns. Above all, there is a sense of traditional philosophical motivation. This remarkable figure will occupy us for a long time to come, and Pears’s study is a very important contribution to the effort to achieve a command of his ideas and what led to them.

Thinking of a parallel project, I hope that we will soon see the publication of the second volume of Brian McGuinness’s fine biography, covering Wittgenstein’s life for the same period in which Pears’s second volume covers the philosophy. McGuinness’s first volume[3] was a subtly-handled evocation of this mesmerising and intolerant genius, of the world from which he came and the other world in which his gift was recognised and allowed to flourish. There is a limit to the extent to which we can expect to understand such a person from within, but McGuinness makes it possible to understand the extraordinary impact he had on others, and what he was like in the full flight of ecstatic and miserable youth. The extension of this scrupulous and deeply-informed treatment into the second half of Wittgenstein’s hyper-conscious life is something to look forward to.

[1] Even other people’s lecture notes are being published, most recently Wittgenstein’s Lectures on Philosophical Psychology 1946-47: notes by P.T. Geach, K.J. Shah and A.C. Jackson, edited by P.T. Geach (Harvester, 348 pp., £35, November 1988, 0 85527 526 X). These are three sets of notes from the same lectures, together with discussion, and give some sense of what Wittgenstein was like in action. They remind one also that while the Tractatus Logico-Philosophicus was a genuine treatise, Wittgenstein’s later writings are more like a distilled version of the bizarre, disconnected activity of philosophical thought and discussion itself – which in most writers results in a highly-structured product unrecognisably different from the process that produced it.

[2] Philosophical Review, July 1968. Pears rightly stresses the importance of these notes, written by Wittgenstein in English between 1934 and 1936, which appear to contain the first appearance of the Private Language Argument.

[3] Wittgenstein, A Life: Young Ludwig (1889-1921) – unsympathetically reviewed in these pages by George Steiner (23 June 1988), who reveals that he has not read the book through (‘One looks in vain for any mention of Fritz Mauthner’), and seems particularly annoyed not to have learned anything about Wittgenstein’s sex life.